1# This regression file covers mistakes made in the 9x9 Computer Go 2# Ladder game between GNU Go 2.7.187 (white) and GoLife I (black) on 3# November 2, 2000. 4 5loadsgf games/golife.sgf 16 61 reg_genmove white 7#? [C7]* 8 9loadsgf games/golife.sgf 26 102 reg_genmove white 11#? [H5] 12 13# In this position the monkey jump can be cut with ko and is smaller 14# than a move at the bottom anyway. 15loadsgf games/golife.sgf 30 163 reg_genmove white 17#? [E3|E2]* 18 19# # H5 is nice but not the only move. Securing the connection to A3 or 20# # expanding around E2 or E3 is also okay. Testcase removed. 21# loadsgf games/golife.sgf 32 22# 4 reg_genmove white 23# #? [H5] 24 25# GNU Go did play this connection. The test is to make sure it never forgets it. 26loadsgf games/golife.sgf 34 275 reg_genmove white 28#? [H5] 29 30# G6 is out of the question here. It's harder to decide between B2 and F4. 31loadsgf games/golife.sgf 38 326 reg_genmove white 33#? [B2|F4]* 34 35# See also test case owl:192 36# F3 does kill black. 37# 38# A B C D E F G H J 39# 9 . . . . . . X O . 9 O=white 40# 8 X X . O . O X X O 8 X=black 41# 7 O X O . . O O O . 7 42# 6 O O O X X X O . O 6 43# 5 . . X . X O . O X 5 44# 4 . . X 3 4 X O O . 4 45# 3 O . X . 2 1 O . . 3 46# 2 . O . X X X X O . 2 47# 1 . . . 7 . 6 5 . . 1 48# A B C D E F G H J 49# 50loadsgf games/golife.sgf 46 517 reg_genmove white 52#? [F3]* 53 54# Connecting the one point ko at G5 is too small. Either C2 or G1 looks 55# largest. 56loadsgf games/golife.sgf 48 578 reg_genmove white 58#? [C2|G1] 59 60loadsgf games/golife.sgf 56 619 reg_genmove white 62#? [C1] 63 64loadsgf games/golife.sgf 60 6510 reg_genmove white 66#? [D7] 67 68